A distributed system usually stores redundant data, which is utilized by its load balancing process. In theory, the more redundant the data, the easier it is to balance the distributed system. However, load balancing a distributed content delivery system is usually a complex task, which may require inter-server synchronization.